Generally, configuration fault localization involves identifying faults in the configuration of a component (or components) that is/are the source of failure given a set of observed failure indications. Configuration faults are harder to detect than on/off failures, as it involves analysis of the parameters that constitute the configuration.
Configuration faults are widely prevalent and continue to present significant challenges, especially given the complexity of current distributed environments. Conventional techniques for localizing configuration faults are heavily reliant on resource dependency information and tightly coupled to event correlation techniques which suffer from limited availability of granular event data. As dedicated resource environments evolve into shared resource models in cloud computing, the complexity of the problem grows considerably due to an increase in system size that creates a considerable increase in events.